DE112005002331T5 - Method, system and apparatus for compiling and using biological knowledge - Google Patents

Abstract

Method for generating new biological knowledge comprising the following steps:
(a) providing a database of biological assertions comprising a plurality of nodes representing biological elements and descriptors characterizing the elements or relationships between the nodes;
(b) obtaining from the database a subset of assertions that satisfy a set of biological criteria that a user has specified to define a selected biological system;
(c) compiling the acquired data to produce a unit that constitutes a biological knowledge bank of assertions that is potentially relevant to said selected biological system; and
(d) transforming said entity to generate new biological knowledge about said biological system.

Description

relative Registrations

These Registration entitles you to use the provisional application No. 60 / 535,352 entitled "Process, System and device for the compilation and use of biological Knowledge ", requested on January 9, 2004, the disclosure of which is incorporated herein by reference.

technical area

The The invention relates to methods, systems and devices for new to identify biological knowledge, in particular procedures, systems and devices to assemble a biological knowledge bank, Methods, systems and devices to subsets of life science To form data and information and into biological models too transform as well as methods, systems and devices for simplification the electronic processing of and the reasoning biological information.

background

The takes in the world today generated amount of biological information dramatically too. It is estimated, that the amount of information is currently every four to five years doubled. With the huge amount of information that processes and needs to be analyzed are traditional methods (or methods) recognizing and understanding the meaning of this information, especially in the field of life sciences, overwhelmed.

Around to an actual understanding To get to a biological system, a researcher needs in the Life sciences combine information from many sources. The Grasping biological systems is by the interdisciplinary nature biosciences. To understand biological systems to be able to can have extensive knowledge in genetics, cell biology, biochemistry, medicine and many other fields of expertise. To biological It may be necessary to understand systems be to link a lot of different information. bioscience Information can be material about basic Chemistry, proteins, cells, tissues and effects on organisms and populations include - the all could interact with each other. These interactions can complex, barely understood or hidden.

It is already trying to electronic models of biological systems to create. This includes the compilation and organization of enormous amounts of data and the Build a system that can use the data to behave like that biological system. Because of the complexity of biology and the sheer volume of data can be the structure of such a system Hundreds of years and cost tens of millions of dollars. These are the ones that provide new insights in the life sciences and seek new knowledge, against the ever more difficult task that correct data from the mountains of information, gathered from completely different sources. Companies that were ready for it Investing funds had so far been unsuccessful in working out of actually usable Models that help researchers significantly improve the bioscience To increase knowledge. Therefore are the current ones developed for the processing and analysis of knowledge Systems for processing and recording life sciences data far from optimal and authoritative new services are needed.

More accurate expressed The industry needs a method to get the huge amounts of knowledge out different life sciences, and with their help to create knowledgeable and meaningful models, the tested and can be queried around new biological relationships, paths, causes and effects and to see other insights efficiently and easily.

Summary the invention

In accordance With the invention it is recognized that the provision of means for the quick and effectively building part-knowledge banks and derived knowledge bases a key is to provide useful and easier-to-use biological knowledge banks, the biological systems can effectively adjust. These specialized knowledge bases can be from a global knowledge base starting from a potentially relevant one Subset of life science relevant data by a user meet specified criteria, A specially designed knowledge base with the set out here is recorded Structure is assembled.

These can be refined in different formats, augmented, examined and presented in various formats, used by human observation and analysis; this can be facilitated by a variety of tools to understand and detect hidden interactions and relationships in biological systems, ie new biological knowledge is created. This, in turn, allows new hypotheses about biological pathways based on new biological knowledge to be set up and allowed Users to confirm or refute a hypothesis by designing and performing biological experiments with biomolecules, cells and animal models, or a clinical trial.

The Invention thus provides a new paradigm, new methods, devices and instruments applied to a global knowledge base can. The tools and procedures allow efficient implementation of Exploration projects in the for the life sciences relevant areas. The invention provides new procedures and tools that allow a knowledge bank to condition for a project or a task both Focusing as well as flexibility to enable. The invention also makes it possible to treat every biological topic no matter how obscure or esoteric, provided there is at least some assertions in a global knowledge base that for this Topic are relevant. Assertions are true statements that are based on an existing object in a system or a true statement about a Object in this system and a literal value or any one Combine it. Every true statement in a knowledge bank or a unit is referred to herein as an assertion.

One Aspect of the present invention is the detection of a subset from data that is necessary or helpful, from a global one Knowledge base or a central database and the rebuilding of a more specialized knowledge bank, specifically for the present Purpose is designed. For this it is important that the structure of the global knowledge bank is designed to be a knowledge bank can capture the relevant relationships between the information the partial knowledge bank. The knowledge bank, or what here for the sake of simplicity, is called a unit a selection and the logical organization of seemingly different ones Data in a coherent Model of any biological system defined by any desired Combination of criteria. These units are microcosms of the global Knowledge base, you can in the area they cover, more detailed and comprehensive than the be a global knowledge bank and easier, more productive and more efficient be used. Units can together they can, they can Augment each other or back to the global knowledge bank added become. As already mentioned, the terms unit and knowledge bank are interchangeable.

One important aspect of the invention is that it is the generation of derived units allows. Derived units are those in which based on logical Inferences from assertions are generated new assertions. Derived units can by drawing conclusions and other algorithms are augmented. The augmentation is through The addition achieved new knowledge that may, but not necessarily part of the original entity or the global one Knowledge bank is. Augmentation involves pulling locks unit and alignment of the unit with external data (e.g. Laboratory data, clinical data, literature data, etc.), but is not limited to this.

The Invention provides methods for building a knowledge base, means for their preparation and instruments for their refinement. In particular, the invention provides methods to provide a biological Building a knowledge base by first creating a database of biological Assertions or means such as a user interface for accessing such a knowledge base provided comprising a plurality of nodes which represent biological elements, and descriptors that describe the elements or relationships between them characterize. A preferred knowledge base is in parallel pending U.S. Patent Application Serial No. 10 / 644,582 of the same assignee the disclosure of which is incorporated herein by reference. Next the method captures a subset of assertions from the knowledge base, which meets a set of biological criteria that a user specifies has to be a selected one to define a biological system. Then the collected data is compiled, by one unit, d. H. a biological knowledge bank of assertions, the for the selected one biological system could be relevant to create.

The invention provides methods for discovering new biological knowledge. The methods include providing a database of biological assertions that includes a plurality of nodes that represent biological elements and descriptors that characterize the elements or relationships between them. The methods also include detecting a subset of assertions that meet a set of biological criteria that a user has specified to define a selected biological system. The methods also include compiling the detected assertions to create a biological knowledge bank of assertions that may be relevant to the selected biological system, and then analyzing the biological knowledge base to discover new biological knowledge. The invention also provides methods for generating new biological knowledge by providing a database of biological assertions comprising a plurality of nodes representing biological elements and descriptors characterizing the elements or relationships between them, and then a To transform a variety of biological assertions to produce a derived knowledge network.

The Invention provides methods for using a biological knowledge bank including the provision of a database of biological assertions, which includes a plurality of nodes, which biological elements represent, and descriptors that describe the elements or relationships between them characterize a variety of biological assertions to transform and thus create a derived knowledge network and to draw new biological knowledge from the unit.

The The invention provides systems to a biological knowledge base together. The systems include a database of biological Assertions in electronic form containing a variety of nodes includes which biological elements represent and descriptors, which characterize the elements or relationships between them. The Systems also include an application for acquiring a subset assertions that meet a number of biological criteria that a user has specified to a selected biological system define, from the database. The systems also still include a knowledge assembler for the compilation of the detected assertions is configured to to create such a biological knowledge bank of assertions that for the selected biological System are potentially relevant. The invention also provides systems ready to build a biological knowledge base including a database assemble biological assertions that are a variety of Have nodes that represent biological elements and descriptors, which characterize the elements or relationships between them, and an application for transforming a variety of biological Assertions to create a derived knowledge network.

The Invention provides computer for compilation of a biological Knowledge bank and to identify new biological knowledge. The computers include means for accessing an electronic database biological assertions that includes a variety of nodes, which represent biological elements, and descriptors that characterize the elements or relationships between them, as well as a user interface for specific biological criteria used by this calculator to build a unit be used, which is a selected one represents biological system. The calculators also include a Computer application for detecting a subset of assertions, meet the user-specified biological criteria the database, and a knowledge assembler, for compiling the captured Assertions is configured to be a biological knowledge base of assertions necessary for the selected biological System are potentially relevant. The invention also provides a Ready for use with a computer readable program support on the computer readable instructions are included to the above apply the methods and systems mentioned.

In In various embodiments, the invention includes methodological Steps, applications, and calculators to draw conclusions from the assertions so that logical inconsistencies in the knowledge base are eliminated can; to draw conclusions from the recorded assertions, thereby new biological knowledge is generated; to get out of the detected assertions conclusions in order to augment the assertions in it by the knowledge bank additional Assertions from the database that meet the selection criteria, adds; or to augment the assertions in it by adding additional Assertions from data sources outside adds the database to the knowledge base.

In In various embodiments, the invention includes methodological Steps, applications, and calculators to draw conclusions from the assertions to augment the assertions in it by adding additional Add assertions that are new in the unit to the knowledge base; by doing to do a path analysis for the knowledge unit performs, to capture one or more paths that are experimental or clinical data relates; by looking at the detected assertions applies a homologous transformation; by entering with the Assertions performs a logical simulation or by giving the unit additional Assertions from data sources outside adds to the database.

In In various embodiments, the invention includes methodological Steps, applications and calculators to get out of the biological assertions to conclude new assertions; to capture a subset of assertions from the database, the user-specified set of biological criteria to set a selected one biological system met; mathematical operations on series of biological assertions to create new assertions; and biological Assertions to create new assertions.

In various embodiments, nodules are enzymes, cofactors, enzyme substrates, enzyme inhibitors, DNA, RNA, transcriptional regulators, DNA activators, DNA repressors, signaling molecules, transmembrane molecules, transport molecules, scavenging molecules, regulatory molecules, hormones, Zy tokines, chemokines, antibodies, structural molecules, metabolites, vitamins, toxins, nutrients, minerals, agonists, antagonists, ligands, receptors or combinations thereof. In other embodiments, nodes are protons, gas molecules, organic molecules, amino acids, peptides, protein regions, proteins, glycoproteins, nucleotides, oligonucleotides, polysaccharides, lipids, glycolipids, or combinations thereof. In further embodiments, nodules include cells, tissues or organs or molecules that may be considered drugs.

In different configurations can the biological information represented by the nodes and assertions experimental data, knowledge from the specialist literature, patient data, data from clinical trials, compliance data, chemical data, medical Include data or hypothetical data. In other embodiments, the biological information true statements about a molecule, a biological structure, a physiological state, a property, a phenotype or represent a biological process.

In different embodiments represent the biological Information a molecule, one biological structure, a physiological state, a property, a phenotype, a biological process, clinical data, medical data, Data about Diseases or intolerances. In various embodiments, the biological information includes a descriptor for Condition, location, amount or substructure of a molecule, biological Structure, physiological state, property, phenotype, biological Process, clinical data, medical data, disease data or Intolerances.

In Various embodiments include that generated by the method new biological knowledge predictions about physiological behavior People, for B. from the analysis of experiments carried out on animals like studies about Medication efficacy and / or drug toxicity or on the Identification of biomarkers based on prognosis, diagnosis, susceptibility for medicines, drug toxicity and the severity or stage of the disease. In some embodiments includes the method of comparing different units, in others Data mapping and in turn the graphical representation all or more parts of the unit so as to understand that Extrapolation, interpolation, and drawing conclusions facilitate.

The Previous and other features and advantages of the present Invention and the invention per se will become apparent from the description, the drawings and claims, the follow, better understandable.

Short description the drawings

In In all views of the drawings, like reference numerals refer continuously on the same parts. The drawings are not necessarily scale, Instead, the weight was quite general on the presentation the basics of the invention. In the following description be different configurations with reference to the following Drawings in which:

1 an overview diagram is that illustrates an illustrative embodiment of the invention.

2A an original network shows and 2 B shows a subset of a network in accordance with an illustrative embodiment of the invention.

3 shows a knowledge unit graph in accordance with an illustrative embodiment of the invention.

4 shows the union of two paths in accordance with an illustrative embodiment of the invention.

5 shows a knowledge unit graph in accordance with an illustrative embodiment of the invention.

6 shows a knowledge unit graph in accordance with an illustrative embodiment of the invention.

7 a transformed network in accordance with an illustrative embodiment of the invention.

8th Figure 3 shows the presentation of a pooled metabolic reaction in accordance with an illustrative embodiment of the invention.

9 shows a derived network in accordance with an illustrative embodiment of the invention.

10 an illustrative example of data mapping in accordance with an embodiment of the invention.

11 Referring to parent causes inference paths, beginning with a change in messenger RNA levels for a particular gene in accordance with an illustrative embodiment of the invention.

12 Figure 12 is a diagram showing the propagation of predicted changes in a forward simulation compared to observed expression changes in accordance with an illustrative embodiment of the invention.

13 Figure 12 is a diagram generated in accordance with an illustrative embodiment of the invention by a backward simulation of nine expression data points, followed by truncation of the graph to show only the consequence chains supporting the primary hypotheses.

14 In accordance with the present invention, an illustrative example of a visualization technique based on a forward simulation that compares predicted results to actual laboratory data is shown.

15 shows a unitized graph in accordance with an illustrative embodiment of the invention.

16 Figure 4 is a graph showing simulation results in accordance with an illustrative embodiment of the invention.

17 Figure 12 shows a plot of time-series expression and mapping of proteometric data to a segment of a known metabolic pathway in accordance with an illustrative embodiment of the invention.

18 Figure 3 shows a diagram indicating a means for summarizing time, dose, or data from other series of data from many experiments around a particular gene or protein, in accordance with an illustrative embodiment of the invention.

19 Figure 4 shows a pie chart that summarizes the correspondence of a hypothesis with observed data, in accordance with an illustrative embodiment of the invention.

20 in accordance with the invention shows an example of an algorithm for use in evaluating a biological model by comparing predicted results to actual results.

21 in accordance with the invention shows an example of an algorithm for identifying a biomarker.

description

Around the current one Implementing this invention will become a global knowledge base or central database structured so that they have a variety of nodes and descriptors includes and copies or transfers these nodes and descriptors can be without losing internal consistency or biological context goes. Nodes are elements of biological systems, both physical as well as functional, and for example, include such things like certain organs, tissues, cells, organelles, cell compartments, Membrane, proteins, DNA, RNA, small molecules, drugs and metabolites one. The descriptors are data entries that make the nodes functional and / or structurally connect to each other (eg, case frames that are "verbs" that the Determine correlation of nodes) and data entries that additional Information with one or both nodes and their interrelations (eg recording the species or organ, where the protein is found, determine the journal in which the data is published became note of tertiary structural information about the treated protein, note that the protein in patients increased with hypertension is, etc.). The global central knowledge bank can be a large amount contain information, and do most of the same for the present Task are not relevant, yet it has a structure that has a Capture potentially relevant assertions allowed on the application of user specified biological criteria based.

Nodes, although not exhaustive, may be biological molecules, including proteins, small molecules, ions, genes, ESTs, RNA, DNA, transcription factors, metabolites, ligands, transmembrane proteins, transport molecules, scavenging molecules, regulatory molecules, hormones, cytokines , Chemokines, histones, antibodies, structural molecules, metabolites, vitamins, toxins, nutrients, minerals, agonists, antagonists, ligands or receptors. The nodules may be drug substances, compounds that are considered candidates for drugs, antisense molecules, RNA, RNA interference, hairpin RNA, double-stranded RNA or chemogenetic or chemoproteomic studies. Chemically, the nodes can be protons, gas molecules, small organic molecules, amino acids, peptides, protein domains, proteins, glycoproteins, nucleotides, oligonucleotides, polysaccharides, lipids or glycolipids. In higher order models, the nodes may be protein complexes, protein-nucleotide complexes such as ribosomes, cell compartments, organelles or membranes. Structurally, they may be various nanostructures such as filaments, intercellular lipid bilayers, cell membranes, lipid rafts, cell adhesion molecules, tissue barriers and semipermeable membranes, collagen structures, mineralized structures or connective tissue. In even higher orders are the nodes Cells, tissues, organs or other anatomical structures. For example, a model of the immune system may include immunoglobulins, cytokines, various leukocytes, bone marrow, thymus, lymph nodes and spleen. For example, in simulating clinical trials, the nodules may be individuals whose clinical prognosis or current symptoms, medications, levels of drug dosage, and clinical endpoints. In simulating epidemiology, for example, the nodes may be individuals, their symptoms, physiological or health characteristics, environmental factors to which they are exposed, substances they are taking, and disease diagnoses. Nodes may also be ions, physiological processes, diseases, disease processes, translocations, reactions, molecular complexes, cell components, cells, anatomical parts, tissues, cell lines, and protein domains.

descriptors can represent biological relationships between nodes and include noncovalent ones Bonding, Adherence, covalent modification, multimolecular interactions (complexes), cleavage a covalent bond, conversion, transport, change of state, Catalysis, activation, stimulation, agonism, antagonism, upregulation, Repression, inhibition, down-regulation, Expression, posttranscriptional modification, posttranslational Modification, internalization, degradation, control, regulation, chemoattraction, phosphorylation, Acetylation, dephosphorylation, deacetylation, transport and transformation, but are not limited to these.

A preferred form of descriptors for use in the invention are, for example, case frames that are out of the representation structure what is an immediate and general application of the Models on a variety of life sciences or others Systems allowed. Case frames are described in detail in the pending US patent application described with the serial number 10 / 644,582 of the same applicant, the disclosure of which is incorporated herein by reference. Descriptors can be quantitative Functions such as differential equations include the possible quantitative ones Represent relationships between node pairs that are used can, to further refine the network. Descriptors can also include qualitative properties that are either not measured or difficult to describe analytically or quantitatively, or due to insufficient Knowledge of a system in general or the property itself impossible described can be.

Of the Process of knowledge gathering can be on dissimilar systems applied, and the result can be summarized into a collection unit which establishes a model. Furthermore may be based on a knowledge unit based on dissimilar systems is to be accessed as a closed model by clicking on the Fragments of the model is accessed decentrally. A model poses a hypothesis explaining the operation of the system, d. H. Based on a simulation, it is able to predict data to produce that to the actual Fit data that serve as eligibility criteria. The hypothesis can be tested by further experiments, they can with others. Models or networks, it can be refined, verified, reproduced, modified, perfected, corrected or with new ones Nodes and new assertions based on manual or computer aided analysis new data is based, expanded and productive as a biological knowledge bank Models of parts of a physiological path or are used Subnetwork in a cell compartment, a cell, an organism, a population or ecology can be linked to a closed model by a or multiple nodes in a model with one or more nodes in connected to another model.

each true statement in a knowledge bank or a unit becomes here referred to as an assertion. Assertions are true statements, that relate to an existing object in a system, to a true statement about an object in this system and a literal value or any one Combination of it. In different embodiments can assertions Knowledge such as RNA, proteomic, metabolic or clinical knowledge Sources such as medical publications, patient data, data from clinical trials, compliance data, chemical data, medical data, hypothetical data or data from biological databases.

Building a unit begins as soon as an individual, through input through a user interface, specifies biological criteria designed to retrieve from the central knowledge base all assertions that are considered potentially relevant to the topic being addressed. The example classes of criteria applied to the central knowledge base to produce the raw unit include assignments, specific networks (e.g., transcriptional control, metabolism), and biological contexts (e.g., species, tissues, developmental stage), but are not limited to these. Additional example classes of criteria include assertions based on relationship descriptors, assertions based on regular-word matching, assertions calculated using forward-chaining algorithms, assertions based on homo based on logic calculations, and all combinations of these criteria, but are not limited to them. Often key words or root words are used, but other criteria are valuable. You can z. For example, assertions based on various structural algorithms, e.g. For example, one may use forward chaining or backward chaining algorithms (eg, capture all linked assertions three or more steps below all serine kinases in mast cells). Different logical operations such as "or", "and" and "not" can be applied to any selection criteria to specify more complex selections, it is the diversity of the series of criteria that can be worked out and the depth of the assertions in the global knowledge base that allow the flexibility inherent in the invention.

assertions, in agreement with the invention in the form of data entries, a number of specific Fulfill criteria, selected are retrieved from the knowledge bank and then in a knowledge bank or unit, which consists of a subset of nodes with interrelationships and descriptors for the examined system is potentially relevant, exists again composed. By this subsets formation becomes a new biological Model created. This model typically includes far less Assertions as the global knowledge base and serves as a starting point on the way, a more useful Create unit with a smaller focus. After that it will be automatic Routines in the software application that created it, and thereby that the person executing the application uses tools transformed and refined. It can be augmented by other information and other information can incorporated therein, including, but not limited to, assertions that come from the literature and from the person giving the data as relevant for the biological system is considered.

Units, which were created by the present invention are usually better as the global knowledge base or the central database, from the They were derived as they typify the actual biology better predict and describe. This achievement of the invention is based on the application of logic during or after compilation the raw data series to augment the initially retrieved data and to improve the resulting structure as noted herein and more convenient to shape. This can be done automatically when creating the unit, for example through programs embedded in computer software or through the use of software instruments the person who performs, selects and controls the application.

A Unity is in many ways with a global central knowledge bank structurally identical, but it is smaller and much stronger on the considered topic or problem focused, computationally easier steerable and either physically or virtually isolated to them tailored to a particular project, and facilitates the Compliance with conditions of limited use or disclosure requirements, possibly imposed by a data source. Besides, a unit often becomes Have the characteristics of an ongoing work, which in the course of the Application changed, improved, tested and corrected. A unit can be anytime or at anytime Step saved in a predictable format and global Knowledge base added again become.

The Creating a valuable unit therefore involves a process of the subset or segmentation that points to a global central knowledge bank applied, followed by data transformations or manipulations, to improve, refine, and / or enhance the first unit created augment to complete the unit and bring unity to the unit To adapt analysis. That's through the implementation of a process like applying logic to the resulting database she with the actual Biology achieved. So can the criteria after all Ask proteins that occur in human muscle cells, and the central database closes perhaps muscle cell proteins found in mice and some of which are not found in human muscle cells, Then, these data are from a unit representing the physiology of human muscle cells examined, removed. A unit can be created by inserting new ones Nodes and relationship descriptors that come from the knowledge base and based on the assumptions set out above (and many other logical assumptions that are possible) are augmented. A unit can be filtered by adding subsets of data, which are based on other biological criteria excludes. The granulation of the system can be refined or coarsened as it is for the upcoming one Analysis is appropriate (which is essential for the ability to make valid extrapolations between To be able to make kinds or generalizations within a kind, since Data series in their granularity vary). A unit can be made more compact and relevant by making detailed knowledge stronger concluding assertions is summarized for an investigation using data analysis algorithms or for the use of generic Analytical tools such as clusters analysis tools are more suitable.

A unit can be updated regularly following the increase in knowledge and the resulting units can be stored, to show the progression of knowledge in this field. A unit can be augmented in several ways, for example, the person who enters the data can add new data from a structured or unstructured database or data from the literature. A unit can also be returned to a central database so that new assertions can be used as raw material to create another unit.

The underlying presentation of the knowledge of a central knowledge bank is designed to detail knowledge without distortion in relation to grasp the use of knowledge. In a network of such complexity Conclusion can be difficult. Therefore, the methods include and systems of the invention provide a flexible framework for knowledge to process gradually and by the application exactly defined Rules and procedures to create derived units. These derived Units are constructed to be consecutive passages of the Reasoning in the units.

units can used to model any biological system, no matter how it is defined and as detailed as desired, and are only determined by the level of knowledge in the field concerned, access to data and (in the case of new data) the time it takes to manage and import, limited. In one embodiment can Units are used to models continuously or at intervals too update as soon as new data is available for collection, and to provide a means to better understand biology. In another embodiment can Units for the representation of biological systems in whole or in Sharing in different formats can be used to help people they can look at and analyze.

units can Also used in various ways is data on biological systems to learn from new biological knowledge (eg superposition different units to detect differences). In different configurations can Units used to: (a) from the analysis of experiments to predict physiological behavior in humans with animals (eg efficacy and toxicity of medicines); (b) finding ideal biomarkers (easily recognizable or easily quantifiable substances in body fluids to make predictions about Presence of a disease, its prognosis, whether the patient is up Drug X will make severity of illness, etc.); or (c) to learn how to segment members of a population, to improve outcomes and complications in clinical trials to avoid.

units can Furthermore used to study biology by using different ones Compare units (eg human with mouse, diseased tissue with healthy, obese physiology under several different nutritional conditions). units can used to study the biology of tissues at different times while the development, progression or cure of a disease to compare, or to the effect of various disorders like the effects of medication or the effect of another environmental impact within any biological system. units can used to map data (eg by importing experimental data the impact of errors on a biological System to one or more components of the system). In further embodiments may Units used to implement logical simulations, to evaluate data series stored in a global central database not included at the time of creation of the original unit are (for example, a hypothesis based on new experimental data based, retesting) to make assumptions about pathways and complex and subtle causal relationships within a biological To recognize systems and to recognize a cause of illness, to understand toxic biochemical mechanisms and toxic reactions predict.

New knowledge can be ascertained by using the units for example with epistemic motors. Epistemic motors are described in detail in co-pending and filed U.S. Patent Application Serial No. 10 / 717,224, the disclosure of which is incorporated herein by reference. Epistemic motors are programmed computers that accept biological data from actual or imagined experiments to study a biological system and use it to generate a network model of protein interactions, gene interactions, and gene-protein interactions, with the data and the previous one Knowledge of the system, thereby deconstructing the biological reality and suggesting verifiable explanations (models) of the operation of natural systems. The motors identify new interactions between biological structures, for example, between biomolecules, which are the substance of life. These new relationships, taken alone or in concert, explain the behavior of the system. For example, they may explain the observed effect of a disorder in the system, identify factors that maintain homeostasis, the mode of action and side effects of Medika explain epidemiological and clinical data more appropriately, reveal the reasons for the success of a species, unveil embryological processes and recognize the mechanisms of diseases. The programs reveal subtle patterns in complex series of data that the human mind can not track without help. The output of the epistemic motor makes it possible to better understand the system under investigation, to hypothesise, to integrate the studied system into other systems, to form more complex and understandable models and to propose new experiments to test the validity of hypotheses.

The Operation of the systems and methods set forth herein may be considered Software on a computer for general purposes are implemented. In some embodiments can be a computer program in one of the higher computer languages like FORTRAN, PASCAL, C, C ++, LISP, JAVA or BASIC. About that In addition, a computer program can be scripted, macro or in others Functions in commercial software embedded are like EXCEL or VISUAL BASIC, to be written. additionally could Software in assembly language, based on a microprocessor implemented, which is resident on a computer, implemented become. For example, could Software can be implemented in assembly language Intel 80 × 86, though It is configured to work on an IBM PC or a PC clone running. Software can be embedded in an article of manufacture, including one Storage medium or computer-readable medium such as a floppy disk, a hard disk, an optical disk, a tape, a PROM, an EPROM or a CD-ROM, but not limited to these.

contraction of units

The Invention allows the creation of knowledge units by means of acquisition from a global central database and then adding new knowledge by incorporating data and other procedures. In An example will be new knowledge of a global central database added in a step-by-step process, its focus on the application lies. First, the global central database becomes general knowledge, added to the global central database (eg. additional Know about Cancer). Second, basic knowledge in the query area of the intended use (eg prostate cancer) from the medical literature collected, belong to including textbooks, scientific writings and reviews of articles. thirdly becomes the special focus of the project (eg androgen independence in prostate cancer) used to provide even more specialized information sources select. This is followed by the use of experimental data to the next step to feed in data collection and knowledge gathering. To the Example can experimental data show which genes and proteins are focused Involved in the field. By entering data from the specialist literature, which relate to genes and proteins can be a subunit which focuses on the specific area of interest.

1 shows in accordance with the invention an illustrative overview of a system. In this diagram, the system becomes 100 used to detect new biological knowledge. In phase 110 A global central knowledge bank is created by entering information (eg scientific data from the literature, public databases and information from the study of publications) into a computer database. In phase 120 Collects a subset of the information in the global central knowledge bank to generate knowledge units based on biological content. The knowledge units are then refined. In phase 130 For example, experimental data (eg, data related to proteins, RNA, metabolic activity, clinical information, etc.) is used to guide the insertion and collection of knowledge. In phase 140 For example, knowledge units can be used in a variety of applications, including, for example, data mapping, focused assembly by application of pathfinding, graphical output, and logical simulation.

Around To create derived units, algorithms can be used become. In some embodiments, algorithms may be used as computer programs be expressed and used to derived units as data objects within a programming framework An example algorithm leads one or more transformations on the existing units, to generate a new unit. Transformations can be for Example can be achieved by any of the following techniques: (a) by selecting assertions from existing units and the selected ones Insert assertions into a new unit under construction; (B) by combining nodes and assertions from existing units and the combined nodes and assertions into a single entity inserts; (c) by performing mathematical set theory operations on the nodes and assertions from existing units and applies the nodes and assertions resulting from these operations into one Unit inserts; (d) by putting together compilation operations on existing units applies to create a unit for further transformations is used, or (e) by using any combination of applies the above-mentioned techniques.

The simplest form of transforming a unit is to create a subset of the unit. For example, a subset of a unit may contain a subset of the nodes and descriptors in the original unit. A subset is essentially the result of a query that selects nodes and assertions based on a set of criteria. These criteria can be defined procedurally, eg. For example, the selection may be the result of an algorithm that iteratively or recursively explores nodes and descriptors embodying the entity. Like in 2A shown, became an original network 200 of knots 210 and descriptors 220 as in 2 B shown transformed to a subset network 205 from knots 210 and descriptors 220 A bindingInput B is an assertion that links a class A of molecular binding processes to a class B of molecular entities (eg, molecule or complex).

In In some embodiments, a unit may take the form of one or more Accept database tables with columns and rows. In these embodiments may be the transformation or subsampling of a global knowledge base to reach a unit for example by selecting rows which represent assertions from a database table, the match the user's selection criteria. It should be understood that a knowledge bank or a unit in the form of a database only one way is where information can be represented in a computer. Information could instead, as a vector, as a multidimensional array of Data, as a linked data structure or in the form of many other suitable structures or configurations represented by data.

One Aspect of an assertion is an assignment. An assignment presents the source of assertion, for example a scientific one Article, a summary (eg Medline or PubMed), a chapter a book, conference proceedings, a personal message or a internal memo. A unit can be selected by selecting descriptors whose assignment meets certain criteria, e.g. B. match By type of attribution source, attribution source name, or date the assignment source. You could For example, select all assertions whose assignment is a node that is, a journal article published in 2001 or later represents.

One Another aspect of an assertion is its biological context. So can Assertions selected which are linked to a particular biological context.

biological Context refers for example to species, tissues, body parts, Cell lines, tumor, disease, sample, virus, organism, stage of development or any combination thereof. Another aspect of a Assertion is their trustworthiness, a measure of Trust that the assertion is truly representative, actual Biology is reflected and reproducible. Assertions can also according to their trustworthiness selected become. It sets a minimum threshold and all assertions that meet or exceed this threshold are selected.

subsets a knowledge bank also by using specifications that form a complex pattern of Assertions between nodes are defined. All rows of nodes and assertions that match the criteria of the pattern the subset. In one embodiment, a search algorithm may be the Filter database to a list of biological entities that according to the predetermined pattern. For example A structure search can be used to get the subset of all Generate reactions that are a product that is phosphorylated, and a molecular complex as a catalyst. This search is all through a molecular complex catalyzed phosphorylated reactions and by a single Protein catalyzed omit phosphorylated reactions.

In another embodiment, subsets may be generated using path search algorithms, including radial path search, short path search and all path search. Radial pathfinding is useful in finding out how a biological entity is functionally or structurally related to another biological entity. For example, if any given cell contains a mutated form of P53, one might be interested in elucidating its effect on parent and minor molecules of the mutant gene product. An algorithm to find this information can begin at a particular node and find all nodes connected to that node by a predetermined number of steps remote from the node. If directionality is important (such as in reactions), the algorithm may be instructed to follow links only in the direction specified by the path search criteria. The radial path search can be performed in several steps. A radial path search in two steps will involve, for example, starting from a node, searching for the nodes directly linked to it and then searching for the nodes directly linked to those nodes. This process can be applied to as many steps as necessary. These Analysis can be used to determine the expected changes and predict when an existing node is disturbed. This analysis can be shown to the user to explain how a change could spread in the knowledge base and thereby determine its actual impact on a biological system. 3 shows an example of the progression of a radial path search in two steps, that of a specified node 300 emanates. In the first step of the search are the connected nodes 310 found. In the second step of the search are the connected nodes 320 found. The result of this radial path search is the combination of all nodes and assertions as in 3 shown. 3. Optionally, a path search may also be configured to follow only certain descriptors, ignore certain nodes that are perhaps ubiquitous or uninformative, or stop searching for new nodes once they encounter certain nodes.

In huge In biological networks, there are usually two paths between two entities. often is the shortest Path the most useful for an analysis. An algorithm for determining the shortest path in a network begins with a radial breadth search of each of the two start nodes perform. Once a common node is found, the path becomes shortest Path between the nodes announced. To the paths between determine multiple nodes, the algorithm discussed above to find the shortest Paths are applied until all paths between the nodes are found Has. This technique starts from every single start node from a radial pathfinder. Then the paths are recorded, which one has followed in each of the radial searches. The result of this Algorithm is the union of all paths from the start nodes to the destination nodes. Because this approach tends to be in relation to the Number of paths and nodes can grow exponentially, the algorithm limited so that he can take a predetermined number of steps follows. So a three-step search will only generate all the paths that exist between the given source nodes by each one Node is performed from a radial search in three steps. The result of this trail algorithm may be, for example, as sorted list of paths be displayed, starting with the shortest or longest, or as an integrated graph.

An integrated graph is generated by summarizing all used paths up to a given length in the case of a radial path search, or by grouping together the series of paths that link any of the source nodes to any of the destination nodes. This can be achieved by combining two paths at the same time until only a single graph is created, containing all the nodes and assertions. An example of combining two pathways includes all common nodes and assertions, as in FIG 4 shown to summarize in a combined path. Because nodes A, B, and D in this diagram are both pathways 410 as well as to trail 420 belong, these nodes are only once in the combined path 430 shown. Node B appears in trail 410 on and node E in trail 420 and they are also combined in the trail 430 shown. 5 shows the result of merging all paths into a single graph, based on a path search between the start node "FXR" (in the upper left corner of the diagram) and a destination node "LDL" (in the lower right corner of the diagram). This type of analysis also allows the study of the implications of observed changes in gene expression studies or changes in concentrations of proteins and metabolites. The analysis will show how the altered entities are related in order to recognize the dependent changes and find changes that are central to the experiment being performed.

The matrix method is another way to study the changes in a knowledge unit graph. Using a list of nodes of interest (eg, statistically significant, highly modulated RNA in an experiment), the nodes are placed in a matrix, with each node being placed with an entry in a column and a row. Then the shortest path is generated for each node pair (redundant pairings are ignored). Then all generated paths are merged as explained above. The matrix method can also be applied by searching not just one path for each cell in the matrix, but by generating multiple paths. This can be done in several ways: (1) you generate all the paths for each pair; (2) generate the top "n" paths starting with the shortest or longest; (3) generate all top "n" paths that are no longer than a predetermined number of steps. The matrix method is also helpful in determining how a number of biological entities are related. 6 shows the result of a matrix method analysis between three nodes, "Acoxl", "LDL" and "FXR", after all the shortest paths between each node pair have been merged.

A derived network is not limited to operations that subdivide, simplify or summarize the initial network. The derivation can embody a theory of knowledge that allows one to conclude new ones To draw facts based on other facts. A prime example of this is the theory that biological mechanisms are conserved and this mechanism depends on gene and protein sequences. Thus, if one mechanism of one species is known, it can be concluded that this mechanism exists in other species if all the genes / proteins involved in this mechanism have the same - homologous - counterparts in the second species. This technique is used to augment knowledge units that focus on a single type of organism. Thus, an entity focused on human biology can be augmented by facts from the mouse's biology by determining which facts about the mouse fulfill the criteria for homology with humans, and then constructing homologous human facts in the entity , The degree of homology is determined by homology point values, which are calculated by comparing the sequences of the genes or proteins. These scores allow you to set commonality limits for a particular purpose - in some embodiments, the homology criteria can be broad to import facts related to other organisms. In other embodiments, the limits can be narrowed down to cover only mechanisms based on the most similar genes and proteins.

A simple example of a derived network is a network of collapsing nodes that need not be distinguished as distinct terms. Thus, the presentation distinguishes the act of "binding" - a process in which entities form a complex - of a "complex" - the result of this process of bonding - this distinction is irritating in many contexts - especially if one is a network in is a graph or groups proteins according to their binding relationships. 7 shows an example network formed of collapsing nodes. In this diagram, the binding of A and B is merged with the node that represents the complex of A and B, and the new node in each case replaces each of the source nodes.

A unit can be transformed by a summary process. A summary begins with the process of forming subsets, where rows of nodes that conform to a specification are selected. Each of these rows can be replaced by a new set of nodes and assertions, usually a simple pattern like a single assertion between two nodes. 8th shows an example of a summary of two reactions, represented as R1 and R2, that share a common metabolite CoA. The assertions in this example are "R1 Reactant M" and "R2 Product M". The pooled connection between Reactions R1 and R2 is presented as the assertion "R1 new relationship R2." A more complex derivative can be used to simplify a network Creating links and substituting a simple link for a complex pattern of relationships between two links can be seen as a process of "summarizing." In this example, a relationship is created between two genes if they meet the following criteria: (1) each gene has a product that acts as an enzyme in a reaction, and (2) in a reaction in which a gene product acts as a catalyst, a product is formed which, in turn, acts as a reactant in another reaction involving the product the other gene acts as a catalyst, the resulting derived network, as in 9 shown links the genes G1 and G2 which adjoin one another in a derived unit. There are many applications for this derived unit. For example, if it includes endorsements for gene expression data, an algorithm may find groups of coregulated genes adjacent to the derived unit. This corresponds to finding reaction pathways that are regulated together.

Transformations of the unit can be performed by mathematical set theory operations. These operations include, for example, forming the intersection, the difference set, and the union set. You can compare units using set theory operations. All set theory operations assume that there are two existing units. Applying the intersection operation for each assertion in a first unit, the same assertion is checked to see if it does not occur in a second unit. If the assertion occurs in a second unit, it is added to the intersection unit. Nodes mentioned in any assertion in the intersection unit are also selected from the first unit and added to the intersection unit. Applying the difference operation for each assertion in a first unit checks the same assertion to see if it does not occur in a second unit. If the assertion does not appear in the second unit, it is added to the difference unit. Nodes mentioned in any assertion in the difference unit are also selected from the first or second units and added to the difference unit. With a union set operation, a union set unit is created. All assertions in a first unit are added to the union unit of measure. Because every assertion in a second unit is now added to the union unit of measure if it is does not exist in the union unit of measure. Nodes mentioned in the union unit of measure are also selected from the first or second units. The union set operation is another way to demonstrate that two or more units can be merged.

One example for a comparison technique in accordance with the invention is the measurement of the evolution of a knowledge unit about a particular Period. This can be achieved by taking a sequence of units takes that are created over time, and the difference between determined for each couple in the sequence. Moreover, in accordance with the invention two or more units are compared. used For example, consider an intersection of two units, with the Both units are not identical, the intersection of the Assertions in the two units determined. The intersection contains the assertions which occur in both units. If one uses for example the Difference quantity of two units, where the two units are not are identical, the difference of the assertions in the two Units determined. The difference set contains the assertions that appear in one of the units, but not in the other. comparisons between units be helpful to similarities and differences biological Explain systems. So could For example, a unit representing a normal system and another unit is a sick system. For a scientist, it would be very enlightening to determine the similarities and differences of two systems.

instruments to draw knowledge from units The present invention may include the analysis of a unit for detection of new biological knowledge. Analysis concludes another algorithmic analysis by computers or Persons performed can be. Algorithms, the path search, homologous reasoning or include simulation-based reasoning, may be new Derive assertions, which are then fed to the unit can, to augment them. Units can be transformed by homology also refined and augmented, assuming that (1) physics and fundamental biochemical properties and Interactions of substances under typical biological conditions remain constant and (2) homologous structures have identical or analogous function to have. contains a global knowledge base, for example, data that Complex C is produced when molecule A interacts with molecule B in a nerve cell collides, then it can be assumed that A + B = C even then applies when A and B collide in a liver cell. When the liver cell assertion The global knowledge base includes node A and node B, but not the descriptor, stating that they are together the complex C, this information can be compiled during compilation Liver unit to be imported into this. Whole cascades of biological activities can be imported into a unit using such logic. Similarly, if a global knowledge base contains the information that a mouse protein M with the mouse receptor R connects to start kidney tubule repair and human biological assays in the knowledge bank one node homologous to mouse protein M and another contain homologous nodes to receptor R, then the interaction and potential Subsequent events from the mouse unit are imported into a unit oriented towards a human biological system. Furthermore can one unit with another that was generated with other criteria combined and then the logical inconsistencies and redundancies be removed for an even better, more complete or narrow focused biological model can be created.

techniques for graphic output

A Knowledge unit can be visually represented as a graph of nodes be created by shortcuts, represent the biological relationships between and among nodes, are connected. These graphs can be examined by a scientist so he can study the biological System understands better and him the discovery of new biological Know about Systems related to the life sciences facilitate becomes. The use of units to provide biologically relevant insights about that How a system behaves can be used in drug discovery and development and in the development of a variety of therapies extremely valuable be. With the techniques described here one can by the use of Units provided by methods and systems of this invention were to arrive at biologically relevant insights. visualization techniques can can also be used to represent knowledge and related data, for the sake of understanding of the user and the recognition of relationships between entities that occur as patterns and accumulations can, to improve.

After having generated graphs using any of the above techniques, you may want to get a better idea of the biological context of the pathways. This can be achieved by starting at each node in the input graph and performing an n-step radial search from each node. This step "stretches" the nodes and the size of the graph. By color coding the nodes to indicate a modulation (as described by experi mental data) one can recognize the relevant changes that are functionally or structurally proximal to the relevant graph, or in other words, to the biological context.

Experimental data can be mapped in one unit by fitting measurements from experiments to the assertions in the unit representing the quantities measured. In this context, mapping means placing visually recognizable clues such as color on a map of trails to indicate which nodes are involved in a process. This can be done, for example, by adjusting nodes that represent gene expression processes to levels of messenger RNA as measured by microareas or other techniques such as RT-PCR. Nodes representing a plethora of proteins can be adapted to data from proteometric measurements. Nodes representing a wealth of chemical substances can be adapted to data from metabolomic measurements. Once mapped, the data can be processed to create simpler node quality features that simplify the use of algorithms for rendering or analysis. For example, convolutional change data may be aggregated based on user-controlled thresholds, and nodes may be annotated with additional quality features such as "up" or "down" to allow the use of simple algorithms for presentation or analysis. Convolutional change data may also be shown by shading, as in 10 where the shading of each expressed gene in the plot (eg, Mat1a, Mat2b, Pemt, Ahcyl1, Bhmt, Bhmt2, Mfmt, Shmt, and Mthdf) indicates its convolutional change in an experiment (i.e., the darker the shading the more greater the convolution change).

logical Simulation can be in accordance also be used with the invention. As a logical simulation One refers to a class of operations that are based on one unit be applied in the observed or hypothetical changes become valid on one or more nodes in the unit and in which the Implications of these changes based on the causal relationships that assertions in the Unit to be represented, spread on the network. A logical simulation can either be forward, then become the impact of the changes from the origin points of the change starting in the downward direction and acting there, or she can be directed backwards be, then the effects of the changes from the origin points the change starting in the direction of upward spread and work there. In any case is a result of a logical simulation a new derivative Network, which consists of the accounts and assertions, which at the Dissemination of the cause or effect were involved. This contains derived network a hypothesis about the system to be examined.

In the case of a reverse simulation based on observed changes in RNA expression levels 11 Inference paths to find overall causes that begin with an observed change in messenger RNA levels for a particular gene. A particular chain of causes could be as follows: Phosphorylation of a transcription factor by a kinase in which the kinase alters the activity of the transcription factor may in turn induce changes in the expression of genes controlled by this transcription factor. This diagram gives a "pseudocode" description of the inferences, which are then performed to find possible causes for each of the observed RNA changes. The assertion types to be examined are not limited to those shown in this diagram. Any assertion in the entity that represents a causal biological link can be included in this type of analysis. Conversely, then each of the possible causes can be examined to find their respective cause. The process may be repeated for any desired number of steps, with nodes in the entity being annotated for their potential role in the causality of the observed changes.

The resulting derived network contains a hypothesis about the potential Causes of the observed data. This can also happen, depending on the procedures for Dissemination of causality, as a hypothesis about the strongest co-responsible and consistent possible causes of observed Data becomes, i. H. a number of possible causes after objective criteria are arranged. This technique is not based on RNA expression data limited, but can be applied to any number of changes which can be expressed as a presentation system, including, but not exclusively, proteometric data, metabolomic data, post-translational Modification data or even reaction time data.

12 is a manually compiled diagram showing the propagation of predicted changes 1210 in a forward simulation showing those with observed expression changes 1220 be compared. This diagram shows the propagation of predicted protein changes 1210 based on an increase in the amount of a compound 1230 over a known path course. In this diagram, balls represent 1240 Proteins represent pairs of adjacent spheres 1250 Represent protein complexes. Thin arrows with T-shaped spikes 1260 indicate inhibitions or causal decreases. Thin arrows with spikes 1270 indicate an activation or a causal increase. Gene expression relationships are indicated by the arrows 1280 displayed. The diagram is intended to illustrate the manner in which changes predicted by a hypothesis can be compared to observed data.

13 is a diagram created by backward simulation of nine observed expression data points 1320 subsequently, the graph was cropped to just the connections 1330 to show that support the primary hypotheses. Every node 1310 in this figure represents either a gene, a protein or a compound. Nine of these nodes 1320 represent changes in the expression of genes in response to a diet rich in polyunsaturated fatty acids. The rest of the diagram is searched by the unit for possible nodes 1310 which, if changed, would produce one or more of the nine changes observed 1320 could explain, and by the removal of nodes 1310 and connections 1330 so that only the best explanations are shown.

derived Networks can as data objects within a programming framework for general Purposes such as a scripting language. Special operators like the methods described above for searching for paths or logical Enable simulation to store, load, and access these data objects. In addition, the Objects using the standard functions of the programming framework to be edited. Because both the input and the output These operations may include the derived networks in larger procedures, in procedures involving biologically significant inferences, and in Method, the theories and techniques for automatic processing biological series and biological knowledge, several Processing steps are combined. By different Criteria can diverse created derived networks and then compared, merged or be edited in another way. A variety of hypotheses how they are contained in these networks can be evaluated, compared and be classified.

One example for a method that consists of techniques mentioned here would like looks like this: (1) Load a bunch of expression data about convolutional change in the unit; (2) lead you one on the convolutional change data based logical reverse simulation by; (3) examine the resulting derived network and choose The most common affected nodes - those which are the most likely to be classified Causes of the observed data are; (4) return for these Row of knots back to the source unit and lead You can use a pathfinding algorithm to discover the derived network finding that is the smallest graph connecting the nodes; and (5) give the resulting derived network as a graph out. Procedures like this example can be considered functions in the programming framework be included and named and reused.

14 FIG. 12 shows a visualization technique that includes an aspect of the present invention based on a forward simulation that compares predicted results to actual laboratory data. This diagram shows the direct subordinate effects of a fault. The far right column shows the expected result of a fault in the system. Each predicted value is compared to the actual values to determine how accurately the predictions explain the laboratory data. A correlation between the predicted result and the actual effect of each treatment can be calculated. In 14 For example, the cells marked with horizontal lines show a significant increase, the cells marked with vertical lines show a significant decrease, the dark cells show no change, and the non-dark cells are not significant. Interferences may include, among other things, the increase or decrease in the concentration of a transcription factor, a small molecule or a biochemical catalyst.

15 shows a unit summary graph that illustrates the connectivity of the underlying unit from which it was generated. It can give a biologist a quick visual overview of the number of assertions, the distribution of various types of assertions in the unit, and the density or degree to which the underlying unit is linked. The visual overview can be used to determine if the underlying unit has sufficient knowledge, enough assortment types, or a sufficient density of assertions in a particular area. Two diagrams representing two different units can be compared side by side to see if one unit contains more knowledge than the other. One way of comparing it would be to compare two diagrams that represent the same knowledge base at two different times in order to visually examine the growth of knowledge. The mechanics for creating the diagram in 15 is as follows: All nodes and assertions in the unit are converted to a diagram by applying an algorithm for building the graph to create a two-dimensional diagram of the unit. The resulting monochrome chart shows the scale of knowledge contained in the unit and can be used as a starting point for other visualizations. The unitized overview graph can be improved by highlighting the assertions containing a particular relationship descriptor with a particular color.

One variation of the unit lookup graph is to generate a graph that shows simulation results, as in 16 shown. This graph can be created by starting with a monochrome unit graph. The results of a simulation are then placed over this diagram. Inference causal chains can be marked by annotating nodes according to their degree of implication. For example, all nodes that are implied and that predict the hypothesis that they will decrease will be annotated by reddening the nodes or by replacing the symbol of the node with another symbol, such as a downward pointing arrow becomes. Other node states can be displayed by appropriate choice of colors or icons. The appearance of the assertions between nodes can also be changed to mark their causal significance. 16 shows the results of a reverse simulation marked dark gray, whereas the rest of the unit is marked light gray. The output as a graph can help a biologist to determine the extent of the effects of a given disorder on the system.

17 Figure 4 shows a visualization of a time series expression and proteometric data mapped to a segment of a known metabolic pathway. In some embodiments, background colors can indicate the quantity and direction of change relative to controls. Each stained cell corresponds to a particular protein and shows either the changes in the expression level of the corresponding gene or the changes in its observed protein abundance. Each column labeled with a time can display data values for a particular experiment in the time series. This method of presentation is intended to show the changes in the modulation of a trail during a series of experiments, in this case a treatment time course. In 17 shading is used to show expression levels over time (ie, the darker the shade, the greater the gene expession).

18 Figure 12 is a diagram indicating a means for pooling time, dose, or other series data from many experiments for a particular gene or protein. A key point is that each horizontal block indicates a specific type of measurement that is appropriate for the measurement. Gene or protein can be assigned. In this example, the protein Anx7 (Mus musculus) is associated with five types of measurement - two are proteometric measurements by 2D gel, three are microarray data sets that yield gene expression measurement data. In this case, the data is expressed as convolutional changes versus controls, but in other cases, it may be desirable to graph absolute values. For each type of measurement, eight convolution changes are displayed as histogram bars. In general, any number of data points up to a practical limit, which depends on the resolution of the presentation medium, can be represented in this way. The bars can be color coded - for example, red to show downward changes, and green to show upward changes - the general tendency for anyone who may need to search hundreds of these rows of data when reviewing a record Better clarify measurement series. The background colors of each bar can also show the significance of the data. Thus, the expression data in the experiment are actually the average of multiple replications of each experiment, therefore each data point can be assigned a statistical significance measurement. In one embodiment, a blue background may indicate the most significant data, p-value <0.01, whereas a purple background may indicate a p-value <0.05. A yellow background could also indicate any higher p-value. This technique allows the user to easily see the details of the data, details that would have been suppressed in more abstract representations such as a network graph, where nodes are simply colored to indicate "up" or "down", but these provisions are off multiple data points.

19 Figure 4 shows a pie chart summarizing the consistency of the changes predicted by hypothesizing the changes observed in a large data set. The data set in this example consists of expression changes due to treatment of liver cells with fenofibrate. The hypothesis is that the changes are due to an increase in the activity of the transcription factor PPARA. The pie chart in 19 shows the following five categories: (1) correct predictions (17%) confirmed by the data; (2) contrary predictions (1%) which are refuted by the data; (3) predictions (27%) that were not observed in the data; (4) Data observations (26%) for which no over there are unanimous predictions; and (5) conflicting predictions (3%) that can not be attributed to a net change in the data.

EXAMPLE 1

Validation algorithm for biological models

An example of an algorithm used to validate a biological model by comparing predicted to actual results is discussed below and in the pseudocode in FIG 20 described. This algorithm assumes that a knowledge base exists that represents a biological system with knowledge-base data from gene expression experiments.

The predicted results can be determined in two stages. First, a backward simulation, as described here, conducted at a knowledge bank to to determine potential causes of gene expression changes. The reverse simulation generates a list of genes and a score for each Gene. The score for each Knot is based on the "votes" he made during the backward simulation had received. At the beginning of the backward simulation become nodes that represent the significantly upregulated genes, assigned positive votes while those who are significantly downregulated get negative votes. During the Simulation will be votes according to a series of rules that correspond to the causal relationships presented in the knowledge base match, copied from node to node. At the end of the simulation will be the score for calculates each node as a set of three digits: the sum the positive votes, the sum of the negative votes and a grand total, which is the sum of positive and negative votes. At this point For example, the set of nodes representing potential causes ("the reasons") may be for the next step used and selected based on the score of each node, or the range of potential causes can be determined manually. On the second level, the votes for all nodes are set to zero and with the selected one A series of causes is carried out as described here, a forward simulation. With the votes are made in the same way, except that they arise from causes potential effects. At the end of the forward simulation Nodes are checked that represent the expression of genes. Those with a positive total score are those of whom the forward simulation predicts that they are up-regulated, and those with a negative total scores are those that predicted down regulated. The results of the forward simulation represent the totality of the predicted results.

The actual Results are divided into two categories based on the gene expression data based. Contains a list up-regulated genes and the other list contains down regulated genes. The genes contained in these lists can by different statistical methods are generated, where absolute Magnitude of the change (eg signal level), relative magnitude of change (eg convolution values), statistical significance etc. become. The genes can but also manually selected become.

After this the predicted and actual Results have been generated, the results for each Gene in the following three cases arranged in tabular form. In the first case, the gene should be predicted be regulated upwards. If the gene is in the current list of the upregulated genes, the "Counter for correct Predictions "increased. If However, the gene is in the current list of down-regulated genes is the counter for opposite Predictions "increased. If however, the gene is not in any of the lists for actual gene expression changes is located, the "counter for not observed predictions "in the second Case, the gene is according to prediction down regulated. If that Gen up in the list of actually up regulated genes, the "counter predictor counter" is increased but the gene is in the list of genes actually down-regulated is located, the "counter for correct Predictions "increased. If however, the gene is not in any of the lists for actual gene expression changes, will the "counter for not observed predictions "in the third Case, there is no prediction for the gene and the "counter for none Net change "is increased.

For each gene that is in one of the lists of genes that are actually regulated up or down, but for which there is no prediction, the "Unexpected Forecast Counter" is incremented, and the five counters are then output as follows (1) "predictive counter", (2) "counter predictor counter", (3) "unobserved predictor counter", (4) "unpredicted predictive counter" and (5) "counter for none net change ". For example, these counters may be in a histogram or a pie chart, as in 19 shown to be depicted. Such pictorial representations provide the scientist with an intuitive means of determining how well the established hypothesis matches the observed data.

EXAMPLE 2

algorithm for the identification of biomarkers

An example of an algorithm for identifying biomarkers in accordance with the invention will be described below and in pseudocode in FIG 21 described. In general, the algorithm deals with data that characterize a protein of interest and assesses it, taking into account a number of factors that would make the protein a suitable biomarker. The algorithm combines measures from a number of sources, assigns them a numeric value, and bundles them together to output a total score that can be used to evaluate a protein. In particular, the proteins that have the highest absolute score have the highest number of similarities to an ideal biomarker. The factors used in this example are gene expression changes by a drug, existing knowledge of the nature of the gene product, and proximity to a known biomarker. The algorithm was applied to data sets derived from an experiment in which gene expression changes in response to a drug were measured over three cell lines of different susceptibility to this drug.

Of the The first step of the biomarker algorithm is to start from one List of known secreted proteins to perform a trail search. at each search step becomes nodes with the minimum distance to one Source node marked, d. H. the number of steps that they are removed from a secreted protein. The second step is to iterate on the list of proteins in the unit. For each Protein on the list is calculated as follows a list of measures: Calculation of directional coefficient and convolution, biomarker and Secretion score, removal of a secreted protein (was calculated in the first step). These dimensions are in one line in an output file written. Convolution calculations relate on data, as folding convolutions across from Controls expressed be, and can be calculated in different ways, for example (1) disease across from Normal state; (2) treatment with drugs versus none Medication; (3) resistance to susceptibility. The direction coefficient is a measure of the rate of change a series of data points. A data series can be, for example at different times or at different dosage levels be measured. A method to calculate the directional coefficient of a Determining series is the application of a linear regression, which results in a straight line best matching the data series fits.

scores for the Directional coefficients are measured by taking gene expression measurements for every Assay that corresponds to the protein, about three Considered cell types. Examinations subject to cross-binding, are ignored. The remaining values are with a reference level where a value of 2 is assigned when the direction coefficient this exceeds a value of 1 if it only reaches half the height of the reference level, or 0 if the direction coefficient is less than half the height of the reference level reached. For negative direction coefficients are assigned negative values. About the Cell lines will look like three patterns, depending on which one is used searched and the examination point values are calculated. For a dose dependent Patterns become the values above adds the cell types. For a resistance pattern becomes the value for the resistant cell line multiplied by 2 and of which the

total subtracted from the values of the two sensitive cell lines. For an efficacy pattern the value of the most sensitive cell line is doubled, the value of partially added to sensitive cell line and of it the value for the resistant Subtracted cell line. The scores of all examinations will be compared and if a sign conflict occurs in a pair, a total score of 0 is reported to indicate a conflict. In all other cases will be the highest or most negative score reported. Calculations for the convolution values are done in the same way.

For the rating biomarkers will record a score of 2 if that Protein is a known biomarker, or a 1 if it is not. In the same way is for secreted proteins recorded a score of 2 when it a (suspected) secreted protein, otherwise a score of 1 is recorded.

The The output file is sorted using an algorithm derived from the Values of measurements calculated a total score. In the current example will only the convolution score is used. Proteins that are the highest absolute Have values (that is, those at the top or bottom of the sorted list), for further evaluation, see if they are good candidates for Biomarkers would be, selected.

The major components of the score of the algorithm are based on gene expression data. For each locus ID, there are values for multiple series of tests that are processed to provide values about directional coefficients and convolutional changes. The measures for each locus ID are calculated by bundling the data from the examinations while looking for characters, which contradict themselves (contradictory signs would result in the score 0). The algorithm can check dose dependency, sensitivity, resistance, and efficacy of the drug, and the score scales differently for each of these traits. For example, if one were looking for a resistance pattern, the directional coefficient would be considered positive if the two resistant cell lines were the same and the sensitive cell line were different, whereas the dose-response relationship would look for a parallel change across all cell lines. The algorithm explained above provides a list which is then sorted by columns, and the genes that are at the top (convolution) are considered to be good biomarkers.

Even though the invention particularly with reference to specific embodiments However, it should be clear that Of those who are particularly skilled in this art, many changes can be made in shape and detail without being affected by Spirit and scope of the invention as defined by the appended claims deviated from. The scope of the invention is therefore through the claims set out in the Annex displayed, and it is intended that all changes made within Significance and scope of the equivalence of claims occur also include.

SUMMARY

Disclosed For example, methods, systems and apparatus for building compilations are biological Knowledge that forms a biological knowledge base, and to select and transmit Bioscience data and information in biological Models based on electronic processing and electronic reasoning to simplify biological information. A subset of data is from a comprehensive knowledge bank or knowledge store extracted to a more specialized child knowledge base or unit designed specifically for the present purposes is. Compounds generated by the invention allow the Selection and rational organization of seemingly different Data in a model of any selected biological system, by any desired biological criteria is defined. From these compilations can be scooped in a simple way and with big ones capacity and efficiency are logically inferred.

Claims (101)

Method to generate new biological knowledge, which includes the following steps: (a) the provision a database of biological assertions containing a variety of Includes nodes that represent biological elements and descriptors, which characterize the elements or relationships between the nodes; (B) obtaining a subset of assertions from the database, that meet a set of biological criteria that a user specifies that she has a selected one define biological system; (c) the compilation of the collected Data to create a unit that is a biological knowledge base of assertions that are potentially relevant to the said selected biological system is; and (d) the transformation the said unit to get new biological knowledge about the to generate said biological system. The method of claim 1, wherein step (d) is the Application of reasoning to said detected assertions to generate new biological knowledge. The method of claim 1, including the additional step the application of reasoning to said detected assertions to eliminate logical contradictions in said unit. The method of claim 1, which includes the additional Step of applying the reasoning to the mentioned detected Assertions for augmenting the assertions therein include the said unit additional Assertions can be added from the mentioned database. The method of claim 1, which includes the additional Step of applying the reasoning to the mentioned detected Assertions for augmenting the assertions therein include the aforementioned unit from the database mentioned additional assertions, which are new, added become. The method of claim 1, which includes the additional Step of applying the trail analysis to said knowledge unit to obtain a trail or multiple trails. The method of claim 1, which includes the additional Step of applying the homologous transformation to said encompassed assertions. The method of claim 1, which includes the additional Step of applying logical simulation to said detected Assertions includes. The method of claim 1, including the additional step of adding additional assertio includes data sources outside of said database. The method of claim 1, wherein said nodes Enzymes, cofactors, enzyme substrates, enzyme inhibitors, DNA, RNA, Transcriptional Regulators, DNA Activators, DNA Repressors, Signal Molecules, Transmembrane Molecules, Transport Molecules, Scavenging Molecules, Regulatory Molecules, Hormones, Cytokines, chemokines, antibodies, Structural molecules, metabolites, Vitamins, toxins, nutrients, Minerals, agonists, antagonists, ligands, receptors or combinations Include from it. The method of claim 1, wherein said nodes Protons, gas molecules, organic molecules, Amino acids, Peptides, protein domains, proteins, Glycoproteins, nucleotides, oligonucleotides, polysaccharides, lipids, Include glycolipids or combinations thereof. The method of claim 1, wherein said nodes Include cells, tissues or organs. The method of claim 1, wherein said nodes molecules the possible candidate for a Drug are included. The method of claim 1, wherein said biological Assertions include information, experimental data, knowledge from specialist literature, patient data, data from clinical trials, Compliance data, chemical data, medical data or hypothetical Represent data. The method of claim 1, wherein said biological Assertions include information that is a molecule, a biological structure, a physiological state, a property, a phenotype or represent a biological process. Manufacturing article with a computer readable program support, on which contain computer readable instructions to the method To carry out according to claim 1. Method to generate new biological knowledge, which includes the following steps: (a) Provide a Database that includes a variety of nodes, the biological Represent elements and case frames representing interrelations between the nodes; (B) Obtaining a subset of case framework structures from the database, that meet a set of biological criteria that a user specifies that she has a selected one define biological system; (c) Compile the collected Case frame structures to create a unit that is a biological Knowledge base of assertions that are potentially relevant to the said are selected biological system. System to assemble a new biological knowledge bank, contain: (a) a database of biological assertions containing a Includes a plurality of nodes that represent biological elements, and descriptors that describe the elements or relationships between the Characterize knots; (b) an application to a subset from assertions from the database to capture a number of fulfill biological criteria, which a user has specified to be a selected biological Define system; and (c) a knowledge assembler that configures is to compile the collected data, so as to be a unit generate a biological knowledge bank of assertions, which is potentially relevant to the named selected biological system is. The system of claim 18, further comprising an application which makes it possible to draw conclusions from the above mentioned assertions in order to Unity logical contradictions remove. The system of claim 18, further comprising an application which makes it possible to draw conclusions from the above mentioned assertions in order to find new biological Generate knowledge. The system of claim 18, further comprising an application which makes it possible to draw conclusions from the mentioned assertions to the assertions in augmenting it by adding additional to the said unit Assertions from said database, said biological criteria fulfill, added become. The system of claim 18, further comprising an application which makes it possible to draw conclusions from the mentioned assertions to the assertions to augment this by adding additional knowledge to the knowledge base Added assertions who are new in the said unit. The system of claim 18, further comprising an application in order to carry out a trail analysis at the mentioned knowledge bank, to a Path history or multiple pathways to capture. The system of claim 18, further comprising an application for homologous transformation to said detected assertions perform. The system of claim 18, further comprising an application to perform a logical simulation on the aforementioned detected assertions perform. The system of claim 18, further comprising an application additional to the said unit Assertions from data sources outside of the to add the mentioned database. The system of claim 18, wherein said nodes are enzymes, Cofactors, enzyme substrates, enzyme inhibitors, DNA, RNA, transcriptional regulators, DNA activators, DNA repressors, signaling molecules, transmembrane molecules, transport molecules, scavenging molecules, regulatory molecules, hormones, Cytokines, chemokines, antibodies, Structural molecules, metabolites, Vitamins, toxins, nutrients, Minerals, agonists, antagonists, ligands, receptors or combinations Include from it. The system of claim 18, wherein said nodes Protons, gas molecules, organic molecules, Amino acids, Peptides, protein domains, Proteins, glycoproteins, nucleotides, oligonucleotides, polysaccharides, Lipids, glycolipids or combinations thereof. The system of claim 18, wherein said nodes are cells, Include tissues or organs. The system of claim 18, wherein said nodes molecules the possible candidate for a drug are, include. The system of claim 18, wherein said biological Assertions include information, experimental data, knowledge from specialist literature, patient data, data from clinical trials, Compliance data, chemical data, medical data or hypothetical Represent data. The system of claim 18, wherein said biological Assertions include information that is a molecule, a biological structure, a physiological state, a property, a phenotype or represent a biological process. Manufacturing article with a computer readable program support, on which contains computer-readable instructions to the system to apply according to claim 18. Computer to compile a new biological knowledge bank With: (a) a means of accessing an electronic database biological assertions that includes a variety of nodes, represent the biological elements, and descriptors that describe the elements or relationships between them characterize; (b) a user-addressable computer application to extract a subset of assertions from the database, that meet a set of biological criteria that a user specifies that she has a selected one define biological system; and (c) a knowledge assembler, which is configured to compile the collected data to create such a unit that is a biological knowledge base of Assertions that are potentially relevant to the said selected biological System are. The computer of claim 30, further comprising a computer application contains which makes it possible to draw conclusions in order to cite the abovementioned assertions Unity logical contradictions to eliminate. The computer of claim 30, further comprising a computer application contains which makes it possible to draw conclusions in order to obtain new biological assays Generate knowledge. The computer of claim 30, further comprising a computer application contains which makes it possible to draw conclusions about the assertions mentioned above in augmenting it by adding additional to the said unit Assertions from said database, called biological Fulfill criteria, added become. The computer of claim 30, further comprising a computer application contains which makes it possible to draw conclusions about the assertions mentioned above to augment this by adding additional knowledge to the knowledge base Added assertions who are new in the said unit. The computer of claim 30, further comprising a computer application contains in order to carry out a path analysis at the mentioned knowledge bank to record a trail or multiple trails. The computer of claim 30, further comprising a computer application contains to perform a homologous transformation on the aforementioned detected assertions perform. The computer of claim 30, further comprising a computer application contains to do a logical simulation with the mentioned assertions perform. The computer of claim 30, further includes a computer application for adding additional assertions from data sources outside of said database to said entity. The computer of claim 30, wherein said nodes Enzymes, cofactors, enzyme substrates, enzyme inhibitors, DNA, RNA, Transcriptional Regulators, DNA Activators, DNA Repressors, Signal Molecules, Transmembrane Molecules, Transport Molecules, Scavenging Molecules, Regulatory Molecules, Hormones, Cytokines, chemokines, antibodies, Structural molecules, metabolites, Vitamins, toxins, nutrients, Minerals, agonists, antagonists, ligands, receptors or combinations include. The computer of claim 30, wherein said nodes Protons, gas molecules, organic molecules, Amino acids, Peptides, protein domains, proteins, Glycoproteins, nucleotides, oligonucleotides, polysaccharides, lipids, Glycolipids or combinations thereof. The computer of claim 30, wherein said nodes Cells, tissues or organs include. The computer of claim 30, wherein said nodes molecules the possible candidate for a Drug are, include. The computer of claim 30, wherein said biological Assertions include information, experimental data, knowledge from specialist literature, patient data, data from clinical trials, Compliance data, chemical data, medical data or hypothetical Represent data. The computer of claim 30, wherein said biological Assertions include information that is a molecule, a biological structure, a physiological state, a property, a phenotype or represent a biological process. Procedures to discover new biological knowledge the following steps include: (a) Provision of a database biological assertions that includes a variety of nodes, represent the biological elements, and descriptors that describe the elements or relationships between the Characterize knots; (b) obtaining a subset of Assertions from the database, which have a number of biological criteria fulfill, which a user has specified to be a selected biological Define system; (c) Compile the collected data to to create a unit that is a biological knowledge base of assertions forms that are potentially relevant to the selected one are biological system; and (d) analysis of said unit, to discover new biological knowledge. The method of claim 49, wherein said biological Knowledge predictions about the physiological behavior in humans from the analysis of Animals carried out Experiments includes. The method of claim 50, wherein said physiological Behavior that includes drug efficacy or toxicity. The method of claim 50, wherein said new one biological knowledge involves identifying a biomarker. The method of claim 49, wherein step (d) is the Repeat the above method with another set of biological Criteria to create a different entity and different ones Includes units to compare. The method of claim 49, wherein step (d) is the Mapping experimental data to a creation unit a graphical output. The method of claim 49, further comprising adding suspected Assertions to the knowledge base includes, by assignment or a lower trusted, so in step (d) more speculative results can be produced. The method of claim 49, wherein step (d) is the Application of path analysis to said unit of knowledge for further detection of one or more pathways. The method of claim 49, wherein step (d) is the Application of algorithms for determining the mechanism includes. The method of claim 49, wherein step (d) is the Application of techniques for presentation includes to knowledge and related To represent data, so the understanding of the user and the Recognition of patterns and accumulations to improve. The method of claim 49, which is the additional Step of applying the reasoning to the mentioned detected Assertions to eliminate logical contradictions in said unit. The method of claim 49, comprising the additional step of applying the inference to said detected assertions to augment the assertions therein by adding Unit additional assertions can be added from the mentioned database. The method of claim 49, which is the additional Step of applying the reasoning to the mentioned detected Assertions for augmenting the assertions therein by said unit additional Assertions from the mentioned database that are new to be added includes. The method of claim 49, which is the additional Step of applying the path analysis to said knowledge unit for obtaining one or more paths. The method of claim 49, which is the additional Step of applying the homologous transformation to said encompassed assertions. The method of claim 49, which is the additional Step of applying the results from a logical simulation to said detected assertions. The method of claim 49, which is the additional Step of adding additional Assertions from data sources outside the said database. The method of claim 49, wherein said nodes Enzymes, cofactors, enzymes, substrates, enzyme inhibitors, DNA, RNA, Transcriptional Regulators, DNA Activators, DNA Repressors, Signal Molecules, Transmembrane Molecules, Transport Molecules, Scavenging Molecules, Regulatory Molecules, Hormones, Cytokines, chemokines, antibodies, Structural molecules, metabolites, Vitamins, toxins, nutrients, Minerals, agonists, antagonists, ligands, receptors or combinations Include from it. The method of claim 49, wherein said Knot protons, gas molecules, organic molecules, Amino acids, Peptides, protein domains, proteins, Glycoproteins, nucleotides, oligonucleotides, polysaccharides, lipids, Include glycolipids or combinations thereof. The method of claim 49, wherein said Nodes include cells, tissues or organs. The method of claim 49, wherein said Node molecules, the possible candidate for a Drug are, include. The method of claim 49, wherein said Biological assertions include information that is experimental Data, knowledge from the medical literature, patient data, data from clinical trials, Compliance data, chemical data, medical data or hypothetical Represent data. The method of claim 49, wherein said biological assertions include information that is a molecule, a biological structure, a physiological state, a property, a phenotype or represent a biological process. Manufacturing article with a computer readable program support, on which contain computer readable instructions to the method to carry out according to claim 49. The method of claim 49, which is the additional Step of applying the setting up of a new on said Knowledge-based hypothesis of a trail and the implementation of a biological experiment with biomolecules, cells and animal models or a clinical trial to confirm the said hypothesis or to refute. Process to generate new biological knowledge, which includes the following steps: (a) the provision a database of biological assertions containing a variety of Includes nodes that represent biological elements and descriptors, which characterize the elements or relationships between the nodes; (B) the transformation of a large number of said biological assertions, to create a unit that includes a derived network. The method of claim 74, wherein said transformation step the closing on new assertions from the mentioned biological assertions includes. The method of claim 75, wherein said new one Assertions include a variety of new nodes, the new biological ones Represent elements, and Descriptors representing the elements or relationships the new node characterize. The method of claim 74, wherein said transformation step the capture of a subset of assertions that are a series of fulfill biological criteria, which a user specified to define a selected biological system, from the database. The method of claim 74, wherein said transformation step applying mathematical set theory operations to said includes biological assertions to create new assertions. The method of claim 78, wherein said new one Assertions include a variety of new nodes, the new biological ones Represent elements, and Descriptors representing the elements or relationships between new nodes characterize. The method of claim 74, wherein said transformation step summarizing said biological assertions for creation includes new assertions. The method of claim 80, wherein said new one Assertions include a variety of new nodes, the new biological ones Represent elements, and Descriptors representing the elements or relationships between new nodes characterize. System to assemble a new biological knowledge bank, With: (a) a database of biological assertions containing a Includes a plurality of nodes that represent biological elements, and descriptors that describe the elements or relationships between the Characterize knots; and (b) an application for transformation a variety of said biological assertions to a single entity to create a derived network. The system of claim 82, wherein said application from the mentioned biological assertions to new assertions closes. The system of claim 83, wherein said new assertions include a variety of new nodes, the new biological elements represent, and Descriptors representing the elements or relationships between new nodes characterize. The system of claim 82, wherein said application comprises The database captures a subset of assertions that are a series meet biological criteria, which a user has specified to a selected biological system define. The system of claim 85, wherein said detected Assertions include a variety of new nodes, the new biological ones Represent elements and descriptors describing the elements or relationships between characterize new nodes. The system of claim 82, wherein said application mathematical set theory operations on named biological assertions applies to create new assertions. The system of claim 87, wherein said new assertions include a variety of new nodes, the new biological elements represent, and Descriptors representing the elements or relationships between new nodes characterize. The system of claim 82, wherein said application summarized biological assertions to new assertions to create. The system of claim 89, wherein said new assertions include a variety of new nodes, the new biological elements represent, and Descriptors representing the elements or relationships between nodes characterize. Method for extracting from a biological knowledge database, which includes: (a) the provision of a database biological assertions that includes a variety of nodes that represent biological elements, and descriptors representing the elements or relationships between nodes characterize; (b) the transformation of a variety of biological assertions to create a unit, comprising a derived network; and (c) the removal from the mentioned unit, to discover new biological knowledge. The method of claim 92, wherein step (c) is the Illustration of experimental data on named unit for creation a graphical output. The method of claim 91, further comprising adding suspected Assertions to the knowledge base includes, by assignment or a lower trusted, so in step (c) more speculative results can be produced. The method of claim 91, wherein step (c) is the Application of path analysis to said unit of knowledge to further create one trail or multiple trails that are refer to experimental or clinical data. The method of claim 91, wherein step (c) is the application of algorithms for determining the mechanism. The method of claim 91, wherein step (c) is the Application of techniques for representation includes to know and related To represent data, so the understanding of the user and the Recognition of patterns and accumulations to improve. The method of claim 91, comprising the additional step of applying the inference to said assertions to augment the assertions therein by adding to said entity from said database additional As ments that are new. The method of claim 91, which includes the additional Step of applying the trail analysis to said unit to create one trail or multiple trails that refer to experimental or clinical data. The method of claim 91, which includes the additional Step of applying homologous transformation to named Assertions includes. The method of claim 91, which includes the additional Step of applying the results from a logical simulation to assertions mentioned. The method of claim 91, which includes the additional Step of adding additional Assertions from data sources outside the said database.